Israel Paid $46.5 Million to Shape What AI Chatbots Say About Gaza, Investigation Finds
Maryam Tariq
An investigation published by Drop Site News on July 28 has revealed that the Israeli government has spent tens of millions of dollars on a campaign specifically designed to shape how artificial intelligence chatbots respond to questions about the war in Gaza. The findings raise pointed questions about information freedom and algorithmic bias, specifically whether the answers millions of people receive from AI tools are being quietly shaped by a paid, undisclosed influence campaign rather than neutral synthesis of available information.
The Contract and the Man Running It
The operation is overseen by Brad Parscale, former campaign manager for President Donald Trump’s 2020 re-election bid, through his digital influence firm Clock Tower X. According to documents cited by Drop Site, the Israeli government’s contract with Clock Tower X is worth 46.5 million dollars in total and currently runs at 4.5 million dollars a month, having tripled from the 1.5 million dollars a month Parscale originally signed for. Since October 2025, the firm has published hundreds of articles across a network of ten websites, including FactSignal.org, Paxpoint.org and Allyvia.org, each built around promoting a specific angle of Israeli government messaging.
The Strategy Behind the Campaign
What makes this campaign unusual is who it was actually built for. Most of the network’s sites average only a few hundred unique visitors a month, meaning they were never designed to reach large human audiences. Instead, disinformation researchers describe the strategy as a form of large language model poisoning, seeding content specifically optimized to be scraped by AI web crawlers and absorbed into chatbot training data and real-time search responses. Hervé Letoqueux, chief executive of digital disinformation research group Check First, said the sites’ moderate tone and use of sourced material makes them particularly effective, since chatbots often try to present multiple viewpoints on a topic, allowing this content to get repeated as one legitimate side of the debate. According to the investigation, the network’s sites were crawled hundreds of times between January and June 2026, including through Common Crawl, a dataset widely used in training major AI systems.
The Specific Content Being Pushed
Drop Site’s investigation documented specific examples of the campaign’s content. One article, titled The Reality Behind Gaza’s Journalists, Terror Ties, Propaganda, and the Laws of War, asserts that a majority of journalists working in Gaza are linked to armed groups. Another set of published articles casts doubt on the killing of Hind Rajab, a five year old Palestinian girl killed by Israeli military fire in 2024, a case that has been widely documented and covered internationally. FactSignal.org presents itself as a fact-checking platform while publishing material discrediting slain journalists and rejecting genocide allegations against Israel, while Paxpoint.org promotes a narrative of Israel as fundamentally peace oriented.
The Disclosure That Gets Missed Anyway
Under the US Foreign Agents Registration Act, every website in the network legally discloses at the bottom of its pages that the material is distributed by Clock Tower X LLC on behalf of the State of Israel. Despite this legal requirement, Drop Site found that when chatbots including Perplexity and Copilot were asked directly about content from these sites, they did not flag the source as part of a state sponsored influence operation, meaning the disclosure exists but isn’t necessarily surfaced to users receiving AI generated answers built from that content.
Why This Story Matters Beyond Gaza
The investigation’s broader significance lies in what it reveals about AI systems as a new frontier for state influence campaigns. Unlike traditional media manipulation aimed at human readers, this operation was explicitly designed around how AI models ingest and synthesize information, meaning its effectiveness depends on infrastructure most users never see directly. Disinformation experts who reviewed Drop Site’s analysis said the campaign appears to be working, meaning people asking AI chatbots about the war in Gaza may increasingly receive answers shaped by this funded campaign without any indication that a government funded operation helped shape that response.
Sources:
· Israel Is Paying Millions to Train AI Chatbots How to Talk About Gaza. It’s Working. — Drop Site News: https://www.dropsitenews.com/p/israel-brad-parscale-ai-chatbots-gaza
· Israel Is Paying Millions to Train AI Chatbots How to Talk About Gaza. It’s Working. — Quincy Institute for Responsible Statecraft: https://quincyinst.org/2026/07/28/israel-is-paying-millions-to-train-ai-chatbots-how-to-talk-about-gaza-its-working/
· Israel pays millions to poison AI chatbots with Gaza propaganda, report finds — Middle East Monitor: https://www.middleeastmonitor.com/20260729-israel-pays-millions-to-poison-ai-chatbots-with-gaza-propaganda-report-finds/
· Israel funded $46.5M campaign to shape AI responses on Gaza war, report — TRT World: https://www.trtworld.com/article/052c16361697
